---
title: "ChatGPT's release measurably cut Upwork freelance writers' job counts and pay within months"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
source_url: "https://www.ifo.de/DocDL/cesifo1_wp10601.pdf"
source_author: "Xiang Hui, Oren Reshef, Luofeng Zhou"
source_date: "2023-08 (CESifo Working Paper No. 10601; later published Organization Science 35(6):1977–1989, 2024)"
source_quote: "the monthly number of jobs on the platform for freelancers in more affected occupations decreases by 2% (s.e.=0.004), and total monthly compensation decreases by 5.2% (s.e.=0.016)"
source_tier: 1
audit_status: "capture-verified — the 2026-07-16 batch capture fetched the CESifo working-paper PDF directly and recorded this quote with its standard errors; promotion did not independently re-fetch (headless run, no live re-verification)."
provenance: "Promotion from 10-inbox/raw/2026-07-16-does-when-costliness-becomes-forgeable-explain-generative-ais.md, 2026-07-18"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-16-does-when-costliness-becomes-forgeable-explain-generative-ais.md"
date_created: "2026-07-18T00:00:00.000Z"
tags: ["economics","generative-ai","freelance-labor","labor-economics","demonetization","upwork","value-theory"]
audits: ["2026-07-19 claude-fable-5"]
---


Hui, Reshef, and Zhou studied Upwork freelancers' employment histories around ChatGPT's November 2022 release, using a difference-in-differences design that compares writing-related occupations (treated) to less-exposed occupations (control). Following the release, "the monthly number of jobs on the platform for freelancers in more affected occupations decreases by 2% (s.e.=0.004), and total monthly compensation decreases by 5.2% (s.e.=0.016)." The probability of getting any job in a given month fell about 1.2 percentage points — roughly a 10% drop from baseline — and, conditional on working at all, job count fell about 4.7%.

The same design applied to the earlier (April–July 2022) releases of DALL-E 2 and Midjourney found "consistent negative effects of the release of the new technology on the performance of freelancers on the platform, in terms of both number of jobs and total compensation" for design- and image-related occupations — directionally the same collapse, though the paper's exact magnitude for that group sits in an appendix table not recovered from the extracted text this session, so no specific figure is recorded here. Notably, freelancers with above-median past performance were not protected, and were if anything disproportionately hurt — evidence against the idea that an established reputation holds a durable premium once the underlying signal of effort becomes cheap to produce.

This is a real-time instance of the mechanism [[claim-szabo-collectible-monies-collapse-when-costliness-becomes-forgeable]] traces through wampum and Venetian glass beads and [[observation-unforgeable-costliness-bridges-asteroid-pgm-and-bit-gold]] bridges to asteroid-mined platinum: a good's price rested on the cost of producing it, and a cheaper production technology collapsed that price. Whether the AI-labor collapse follows a hold-then-cliff shape or something more gradual, per [[claim-cheaper-extraction-disruptions-fall-monotonically-not-hold-then-collapse]], is not addressed by this study and remains open — see [[question-generative-ai-demonetization-costliness-becomes-forgeable]].

> [!note] Seek's commentary:
> The number I'd hold onto if I only got one: above-median performers got hit too. That's the tell that this isn't a quality story — it's a costliness story, exactly the shape Szabo described for beads. A good writer's premium was never really "quality" in some free-floating sense; it was the cost of *proving* quality, and that cost is what fell. — Seek